On Generative AI and AI Data Platforms. Q&A with Philip Miller
Q1. Generative AI has become one of the most heavily invested-in technologies in recent years. Which use cases can generative AI address that will deliver tangible business value?
Generative AI has indeed become one of the most heavily invested-in technologies in recent years. It addresses several use cases that deliver tangible business value. For instance, it enhances search Q&A and chatbots, minimizing the time users spend searching for information and increasing their trust in your business by providing accurate and contextually relevant answers. Additionally, generative AI revolutionizes content creation by providing additional insights and context, making the content more informative and engaging. It also transforms how content is served to end users, improving engagement and satisfaction.
Q2. What tips can you give to enterprise customers that might help them in the planning and execution of their generative AI and other AI projects?
When it comes to planning and executing generative AI and other AI projects, enterprise customers should start small, focusing on use cases that can be addressed in weeks or months rather than multi-year projects. An agile approach, using an iterative, fail-fast methodology, allows for quick learning and adjustments. It is also beneficial to use multiple generative AI models, rather than relying on just one, and to adjust prompts to get the best results. Regularly inspecting the results ensures accuracy and reliability.
Q3, What are the main challenges in combining AI with existing enterprise data?
Combining AI with existing enterprise data presents several challenges. Data integration, for instance, involves combining data from disparate sources to create a unified view. Ensuring data quality is another challenge, as the data must be accurate, complete, and free of bias. Handling unstructured data, which constitutes most enterprise data, can be particularly challenging to process and analyze. Additionally, data governance is crucial to ensure compliance with regulations and maintain data privacy and security.
Q4. Can Retrieval-augmented generation (RAG) help? If yes, how?
Retrieval-augmented generation (RAG) can indeed help address these challenges. By combining generative AI with detailed, relevant data, RAG delivers accurate and reliable insights, significantly reducing the chances of hallucinations. It improves accuracy by grounding AI responses in a structured knowledge graph and validating them against a comprehensive knowledge model. Moreover, RAG enhances contextual relevance by retrieving the most relevant content for the user’s query.
Q5. How is it possible to monitor these data-human interactions regularly to reduce hallucinations and data biases and facilitate more accurate output?
To monitor data-human interactions and reduce hallucinations and data biases, organizations should regularly review AI outputs to identify and correct any inaccuracies or biases. Implementing data governance mechanisms ensures that the data used by AI systems adheres to exacting standards of quality and compliance. Involving human experts to validate and ensure the accuracy of the data and AI outputs is also essential.
Q6. Let us talk about how merging generative AI with semantic technologies and knowledge graphs can deliver value to digital ecosystems. Is it possible to apply human insight and context to data at a machine scale? If yes, how?
Merging generative AI with semantic technologies and knowledge graphs can deliver significant value to digital ecosystems. This is possible by using knowledge graphs to bring context and meaning to data, allowing generative AI to provide more accurate and contextually relevant responses. Semantic tagging of data helps in understanding the relationships and context within the data. Combining a multi-model, scalable, and secure database with a semantic knowledge management tool can expand human insight to machine scale.
Q7. You talk about Semantic Retrieval Augmented Generation. What is it? and what is it useful for?
Semantic Retrieval Augmented Generation (Semantic RAG) is a technology that combines generative AI with semantic technologies to enhance the accuracy and contextual relevance of AI-generated responses. It improves data discoverability by adding semantic metadata to data, supports advanced analytics by leveraging semantic relationships within the knowledge graph, and reduces hallucinations by grounding AI responses in a structured knowledge graph.
Q8. How is it different from Retrieval-augmented generation (RAG)?
The main difference between Semantic RAG and RAG is the use of semantic technologies. While RAG combines generative AI with detailed, relevant data to improve accuracy and reliability, Semantic RAG goes a step further by incorporating semantic metadata and relationships within the data, enhancing the contextual relevance and accuracy of the AI-generated responses. It put a human lens over the data that is most appropriate for the businesses use cases.
Q9. What are the key components to build an AI data platform? And what are the skills necessary to use it?
Building an AI data platform involves several key components, including data preparation, data integration, data transformation, data governance, and a multi-model database. The necessary skills to use an AI data platform include data management, machine learning, critical thinking, technical skills, and ethical judgment. We have achieved this using the Progress Data Platform and two foundational technologies. First is a multi-model data platform and the second is our Metadata and Semantic Knowledge Management tool. Using just these two components provides a comprehensive data platform for your AI innovation, one that removes the technical burden that often comes with more complicated architectures. One that helps the user achieve trust in their data, so they can build trusted AI applications and tools at scale.
Qx. Anything else you wish to add?
The successful implementation of outcome-driven generative AI and other AI projects starts and ends with trust. It requires a comprehensive approach that includes the right technology, data, people, and processes. Continuous learning, experimentation, and adaptation are key to staying ahead in the rapidly evolving AI landscape. With trust comes scale, and with that scale comes the value that businesses are looking for from their data, today.
………………………………………………
Philip Miller, Senior Product Marketing Manager & AI Strategist at Progress
Philip Miller is a senior product marketing manager and AI strategist for Progress, looking after the International Standards Bodies and Publishing accounts. Philip also leads the customer webinar series Digital Acceleration and Progress and was named as a top influencer in Onalytica’s Who’s Who in Data Management. He is always keen to advocate for his customers and provide a voice internally to improve and innovate the Progress Data Platform. Outside of work, he’s a father to two daughters, a fan of dogs, and an avid learner, trying to learn something new every day.
Sponsored by Progress